A5045963637- Statistical Methods and Inference
- Statistical Methods in Clinical Trials
- Advanced Causal Inference Techniques
- Statistical Methods and Bayesian Inference
- Cystic Fibrosis Research Advances
- Gene expression and cancer classification
- Bayesian Methods and Mixture Models
- Diabetes Management and Research
- Health Systems, Economic Evaluations, Quality of Life
- Bioinformatics and Genomic Networks
- Advanced Statistical Methods and Models
- Neonatal Respiratory Health Research
- Machine Learning in Healthcare
- Neural Networks and Applications
- Face and Expression Recognition
- Genetic Associations and Epidemiology
- Statistical Distribution Estimation and Applications
- Tracheal and airway disorders
- Diabetes and associated disorders
- Optimal Experimental Design Methods
- Fault Detection and Control Systems
- Computational Drug Discovery Methods
- Diabetes Treatment and Management
- Pancreatic function and diabetes
- Diet and metabolism studies
University of North Carolina at Chapel Hill
2016-2025
University of North Carolina Health Care
2023-2024
Stanford University
2023
The University of Texas Health Science Center at Houston
2014-2023
Statistical and Applied Mathematical Sciences Institute
2020
Biology of Infection
2018
Engineering Associates (United States)
2018
University of Michigan
2014-2018
North Carolina State University
2014-2015
John Wiley & Sons (United Kingdom)
2015
Despite its relative frequency among autosomal recessive diseases and the availability of sweat test, cystic fibrosis (CF) has been difficult to diagnose in early childhood, delays can lead severe malnutrition, lung disease, or even death. The Wisconsin CF Neonatal Screening Project was designed as a randomized clinical trial assess benefits risks diagnosis through screening. In addition, incidence determined, validity our randomization method assessed by comparing 16 demographic variables.
Abstract There is increasing interest in discovering individualized treatment rules (ITRs) for patients who have heterogeneous responses to treatment. In particular, one aims find an optimal ITR that a deterministic function of patient-specific characteristics maximizing expected clinical outcome. this article, we first show estimating such rule equivalent classification problem where each subject weighted proportional his or her We then propose outcome learning approach based on the support...
Prognosis is of critical interest in breast cancer research. Biomedical studies suggest that genomic measurements may have independent predictive power for prognosis. Gene profiling been conducted to search measurements. Genes the inherent pathway structure, where pathways are composed multiple genes with coordinated functions. The goal this study identify gene Since our fundamentally different from existing studies, a new analysis method proposed. advances beyond alternatives along...
As part of the ongoing Wisconsin Cystic Fibrosis (CF) Neonatal Screening Project, we had unique opportunity to study longitudinal relationship between Pseudomonas aeruginosa (Pa) acquisition and infection developing lung disease in children with CF. The primary objective was determine whether Pa associated a measurable change progression disease. Two outcome measures were used 56 patients who diagnosed through newborn screening: 1) additive chest radiograph score (WCXR), based on average...
Many patients with cystic fibrosis are malnourished at the time of diagnosis. Whether newborn screening and early treatment may prevent development a nutritional deficiency is not known.We compared status identified by neonatal or standard diagnostic methods. A total 650,341 infants were screened measuring immunoreactive trypsinogen on dried blood spots (from April 1985 through June 1991) combining test DNA analysis July 1991 1994). Of 325,171 assigned to an early-diagnosis group, was...
Precision medicine seeks to maximize the quality of health care by individualizing health-care process uniquely evolving status each patient. This endeavor spans a broad range scientific areas including drug discovery, genetics/genomics, communication, and causal inference, all in support evidence-based, i.e., data-driven, decision making. is formalized as treatment regime that comprises sequence rules, one per point, which map up-to-date patient information recommended action. The potential...
Dynamic treatment regimes (DTRs) are sequential decision rules for individual patients that can adapt over time to an evolving illness. The goal is accommodate heterogeneity among and find the DTR which will produce best long term outcome if implemented. We introduce two new statistical learning methods estimating optimal DTR, termed backward weighted (BOWL), simultaneous (SOWL). These approaches convert individualized selection into either or classification problem, thus be applied by...
Personalized medicine has received increasing attention among statisticians, computer scientists, and clinical practitioners. A major component of personalized is the estimation individualized treatment rules (ITRs). Recently, Zhao et al. proposed outcome weighted learning (OWL) to construct ITRs that directly optimize outcome. Although OWL opens door introducing machine techniques optimal regimes, it still some problems in performance. (1) The estimated ITR affected by a simple shift (2)...
Abstract Forest-based methods have recently gained in popularity for non-parametric treatment effect estimation. Building on this line of work, we introduce causal survival forests, which can be used to estimate heterogeneous effects and observational setting where outcomes may right-censored. Our approach relies orthogonal estimating equations robustly adjust both censoring selection under unconfoundedness. In our experiments, find perform well relative a number baselines.
Abstract We develop reinforcement learning trials for discovering individualized treatment regimens life‐threatening diseases such as cancer. A temporal‐difference method called Q‐learning is utilized that involves an optimal policy from a single training set of finite longitudinal patient trajectories. Approximating the Q‐function with time‐indexed parameters can be achieved by using support vector regression or extremely randomized trees. Within this framework, we demonstrate procedure...
Summary Typical regimens for advanced metastatic stage IIIB/IV nonsmall cell lung cancer (NSCLC) consist of multiple lines treatment. We present an adaptive reinforcement learning approach to discover optimal individualized treatment from a specially designed clinical trial (a “clinical trial”) experimental patients with NSCLC who have not been treated previously systemic therapy. In addition the complexity problem selecting compounds first- and second-line treatments based on prognostic...
Introduction Computerized tomography (CT) scanning shows promise as an outcome surrogate for cystic fibrosis (CF) lung disease progression. The scoring system used to convert the CT image numeric data is essential determinant of performance scanning. Methods Three radiologists independently scored 16 high-resolution scans performed on children in Wisconsin CF Neonatal Screening Project. test were selected provide a broad range severity. provided subscores presence and severity 5 findings...
In the Comparison of Medical Therapy, Pacing and Defibrillation in Heart Failure (COMPANION) trial, 1520 patients with advanced heart failure were assigned a 1:2:2 ratio to optimal pharmacological therapy or plus cardiac resynchronization (CRT-P) CRT defibrillator (CRT-D). Use CRT-P CRT-D was associated significant reduction combined risk death all-cause hospitalizations. Because mortality also significantly reduced (optimal versus only), an assessment true hospitalization rates must...
In this paper, we introduce a new type of tree-based method, reinforcement learning trees (RLT), which exhibits significantly improved performance over traditional methods such as random forests (Breiman, 2001) under high-dimensional settings. The innovations are three-fold. First, the method implements at each selection splitting variable during tree construction processes. By on that brings greatest future improvement in later splits, rather than choosing one with largest marginal effect...
The vision for precision medicine is to use individual patient characteristics inform a personalized treatment plan that leads the best possible health-care each patient. Mobile technologies have an important role play in this as they offer means monitor patient's health status real-time and subsequently deliver interventions if, when, dose are needed. Dynamic regimes formalize individualized plans sequences of decision rules, one per stage clinical intervention, map current information...
This paper proposes a novel deep learning architecture involving combinations of Convolutional Neural Networks (CNN) layers and Recurrent neural networks (RNN) that can be used to perform segmentation classification 5 cardiac rhythms based on ECG recordings. The algorithm is developed in sequence setting where the input five second signal sliding windows output rhythm labels. processes as both spectrograms well heartbeats' waveform. Additionally, we are able train model presence label noise....
Accurate assessment of gestational age (GA) is essential to good pregnancy care but often requires ultrasonography, which may not be available in low-resource settings. This study developed a deep learning artificial intelligence (AI) model estimate GA from blind ultrasonography sweeps and incorporated it into the software low-cost, battery-powered device.
The CDK4/6 inhibitor palbociclib blocks cell cycle progression in Estrogen receptor-positive, human epidermal growth factor 2 receptor-negative (ER+/HER2-) breast tumor cells. Despite the drug's success improving patient outcomes, a small percentage of cells continues to divide presence palbociclib-a phenomenon we refer as fractional resistance. It is critical understand cellular mechanisms underlying resistance because precise resistant tissue strong predictor clinical outcomes. Here,...
It is uncertain whether the growth impairment that occurs in children during long-term treatment with glucocorticoids persists after medication discontinued and ultimately affects adult height.
Patients with cystic fibrosis (CF) are susceptible to lower respiratory tract infections Pseudomonas aeruginosa and typically acquire this organism in early childhood. Once P infection is established, eradication may be impossible, progressive lung disease often aggravates morbidity mortality risks. The ability diagnose CF by genetic testing at birth makes it possible determine the temporal sequence of events that result aeruginosa-associated pulmonary infections.To evaluate longitudinal...